Zortix matched this thesis to Anthropic’s public holders (AMZN, GOOGL, NVDA, MSFT) as one way public-market investors are exposed to it. Stakes as disclosed or reported. Not a recommendation.
Where these come from — Anthropic’s exposure, as disclosed or reported
AMZNAmazon — $8.0bn of convertible notes (Q3 2023–Q4 2025), $5.0bn Series G and $5.0bn Series H nonvoting preferred (Q2 2026), plus a $20bn facility with $15bn undrawn; notes convert to nonvoting common at an IPO, subject to an ownership cap. · as of 30 Jun 2026 · Amazon 10-Q for the quarter ended 30 June 2026 (filed 2026-07-31)
GOOGLAlphabet — about 14% of the equity, non-voting, contractually capped at 15% (reported). Alphabet's own filings disclose only aggregate non-marketable equity securities. · as of Apr 2026 · Court filings as reported by The New York Times
NVDANvidia — commitment to invest up to $10bn (announced November 2025). · as of Nov 2025 · Nvidia 10-Q filed 2025-11-19
MSFTMicrosoft — commitment to invest up to $5bn (announced November 2025). · as of Nov 2025 · Company announcement, November 2025
AGIXKraneShares Public-Private AI & Technology ETF — reported 2.8% of the fund · as of 29 Jul 2026 · Fund holdings as reported
BAIiShares A.I. Innovation and Tech Active ETF — reported 0.42% of the fund · as of 30 Apr 2026 · Fund holdings as reported
Open source AI models reach performance parity
The thesis argued is that open-source AI models now perform 95% of standard tasks, driving commoditization at the base model layer and leaving specialized value to proprietary platforms or training data owners.
Keep reading this one
You've read the thesis and who argued it. A free account opens the argument, what validates it, the risks the show raised, and the moment in the episode where it was said — 3 ideas in full a day, no card.